In this workshop, you’ll learn how to train, accelerate, and optimize a defect detection classifier. We’ll start by exploring the key challenges around industrial inspection, problem formulation, and data curation, exploration, and formatting. Then, you’ll learn about the fundamentals of transfer learning, online augmentation, modeling, and fine-tuning. By the end of the workshop, you’ll be familiar with the key concepts of optimized inference, performance assessment, and interpretation of deep learning models.

Learning Objectives

By participating in this workshop, you’ll learn how to:
  • Formulate an industrial inspection case study and curate datasets generated by automated optical inspection (AOI) machines
  • Deal with the logistics and challenges of data handling in an industrial inspection workflow
  • Extract meaningful insights from our dataset using pandas DataFrame and NumPy library
  • Apply transfer learning to a deep learning classification model (Inception v3) 
  • Fine-tune the deep learning model and set up evaluation metrics
  • Optimize the trained Inception v3 model on an NVIDIA V100 Tensor Core GPU using NVIDIA® TensorRT 5
  • Experiment with FP16 half-precision fast inferencing with the V100’s Tensor Cores

Download workshop datasheet (PDF 83.3 KB)

Workshop Outline

(15 mins)
  • Meet the instructor.
  • Create an account at
Understanding Key Concepts
(120 mins)
  • Learn about the key concepts of visual inspection.
  • Understand the problem formulation and data curation.
Break (60 mins)
Transfer Learning and Modeling
(120 mins)
  • Learn how to train and verify deep learning models, based on transfer learning procedure.
  • Get hands-one experience with online augmentation while training to save disk storage of datasets.
  • Take a deeper dive into the nuances of fine-tuning your model.
Break (15 mins)
Understanding Inference and Interpreting Your Results
(120 mins)
  • Focus on production deployment and optimization.
  • Learn how to freeze the trained deep learning model and optimize it using TensorRT.
  • Compare the performance of the optimized model against the original TensorFlow-GPU model and measure the improvement.
Final Review
(15 mins)
  • Review key learnings and wrap up questions.
  • Complete the assessment to earn a certificate.
  • Take the workshop survey.

Workshop Details

Duration: 8 hours

Price: Contact us for pricing.  

Prerequisites: Experience with Python and convolutional neural networks (CNNs)

Technologies: TensorFlow, NVIDIA TensorRT, Keras

Certificate: Upon successful completion of the assessment, participants will receive an NVIDIA DLI certificate to recognize their subject matter competency and support professional career growth.

Hardware Requirements: Desktop or laptop computer capable of running the latest version of Chrome or Firefox. Each participant will be provided with dedicated access to a fully configured, GPU-accelerated server in the cloud.

Languages: English, Simplified Chinese, Traditional Chinese

Upcoming Workshops

If your organization is interested in boosting and developing key skills in AI, accelerated data science, or accelerated computing, you can request instructor-led training from the NVIDIA DLI.